OLLO S5X Headphone Review

[Mulder]

If I were to filter it would be 250Hz (Q=0.5 to 1.0) -2 to -3dB but don't think it is essential as it removes some of the warmth/fullness.
Treble needs to be lowered but this is better done with felt/toiletpaper than EQ.

I tried this EQ setting today. 250 Hz; Q=0.5 and -2 db. (Thanks to rainy weather ) Observe: I used this EQ settings in combination with @solderdude’s suggested felt mod. My conclusion: If there is any distortion/resonances to reduce, then I can’t tell if this makes a difference. The only difference I noticed was that the bass got weaker. I listened to some well recorded bass heavy rock and the difference was clearly weaker bass. The difference was obvious. With aucoustic jazz (Miled Davis ”Kind of Blue”) the bass was not that full as without EQ. It was like the bass was thinner. However, If I had not done an A/B test I am not sure if I had noticed the difference. The difference was not that obvious as in the bass heavy rock example. Can’t say I noticed any imrovements related to what could be less resonances. But then again, I avoid high volume levels. If I was going to use EQ I clearly prefer this EQ settings compared to amirms bass modification, as I think the later compleatly changes this headphones character. This modifikation mostly preserves the overal character.

 

[Mulder]

Probably best to just ignore him. When he can't win an argument on a technical front he resorts to snark, ad hominem and appeals to bogus authority (e.g. himself) as we've seen here. In this instance it all stems from his clearly pseudoscientific personal bias against digital EQ, as evidenced in the 'article' he linked previously:

Hilariously, despite lambasting you for not having his extensive hands-on toilet paper experience, his digital EQ experience is by his own admission decidedly...lacking:


Here's what actual professionals and authorities in the field have to say on the topic. First an acoustic engineer who worked/works at Harman (oh and not only a trained listener but a Harman Golden Ear no less), GRAS & Audio Precision:

Is CSD (Cumulative Spectral Decay) really important?

Frequency response is always under suspicion because a lot of audiophiles claim only single one curve can not describe the whole story especially for the time domain. CSD (Cumulative Spectral Decay) contain an extra dimension——time.

www.linkedin.com

(I've only ever seen one headphone that exhibits audible non-minimum phase behaviour shown by significant excess group delay, the Monoprice M1060. It's exceedingly rare.)

Dr. Floyd Toole (my emphasis):



It follows from this last quote that damping would likely in fact result in a more audible resonance than using digital EQ (and potentially even more audible than without, degrading, not improving the stock sound), the latter able to reduce the resonance peak without reducing the Q. Yet another win for EQ vs damping to go along with automatically reducing distortion when bringing down peaks as already mentioned.

And finally, a specific headphone example (source). Here are CSD (waterfall) plots of the HD800 first without, and then with its 6 kHz resonance peak digitally EQed out:
View attachment 302766
View attachment 302767
Oh look, not only has the frequency response ('back' red curve) peak at 6 kHz been removed by this digital EQ, but so has the ringing from it in the time domain. (Of course this all makes sense when it's understood that these domains are mathematically connected via the Fourier transform). No need to take a trip to the toilet for that precious acoustic dumping damping paper after all. Hopefully there'll be no more talking sh*t now that's been cleared up.

I can't find anything in the quoted text that specifically and as a prop states that digital EQ must be used, ie applied in the digital domain. (The text refers to tests that Floyd Toole did 50 years ago, i.e. around 1970. At that time there were not even digital signal sources like today.)

 

[IAtaman]

We are talking about gadgets that are used to listen to music, a non-critical and fully optional aspect of human existence yet look at all this telenovela level drama. One wonders what do people on forums in which they discuss lipid metabolism or cardiovascular health do when they have disagreements, punch each other on the face or hack in to each others' home security systems or something? Or is it just the audio hobby that creates disproportional drama? Anyway.

@GaryH, the link you shared was good read thanks. I got a question if I may please. I understand for transfer function to be a function of frequency, the amplitude response should be linear. But in the case of a resonance, amplitude response is not linear at the resonant frequency, isn't it?

 

[IAtaman]

I tried this EQ setting today. 250 Hz; Q=0.5 and -2 db. (Thanks to rainy weather ) Observe: I used this EQ settings in combination with @solderdude’s suggested felt mod. My conclusion: If there is any distortion/resonances to reduce, then I can’t tell if this makes a difference. The only difference I noticed was that the bass got weaker. I listened to some well recorded bass heavy rock and the difference was clearly weaker bass. The difference was obvious. With aucoustic jazz (Miled Davis ”Kind of Blue”) the bass was not that full as without EQ. It was like the bass was thinner. However, If I had not done an A/B test I am not sure if I had noticed the difference. The difference was not that obvious as in the bass heavy rock example. Can’t say I noticed any imrovements related to what could be less resonances. But then again, I avoid high volume levels. If I was going to use EQ I clearly prefer this EQ settings compared to amirms bass modification, as I think the later compleatly changes this headphones character. This modifikation mostly preserves the overal character.

I think that is way too wide of a filter to isolate the effect of such narrow distortion, if any. Q=0.5 is 2.5 half octaves or so if I am not mistaken - no surprise it made the whole upper bass / lower mids weaker. From what you are describing, I think it is unlikely you are experiecing any distortion effects, but If you wanted to check for that, I think you will need a much narrower filter of at least of Q=3.

 

[Mulder]

I think that is way too wide of a filter to isolate the effect of such narrow distortion, if any. Q=0.5 is 2.5 half octaves or so if I am not mistaken - no surprise it made the whole upper bass / lower mids weaker. From what you are describing, I think it is unlikely you are experiecing any distortion effects, but If you wanted to check for that, I think you will need a much narrower filter of at least of Q=3.

I just tested with Q=3. I have only briefly listened to the same music as before. The difference is hardly or not at all noticeable in terms of the bass reproduction. Just like before, I have no impression that there are resonances that are taken care of. No increased clarity or anything like that. But I didn't expect it either. Especially not at the modest volumes I listen to, even if I turned the volume up a little extra just for the sake of it.

 

[Mulder]

I've got plenty of time today and spent several hours listening to the SX5 (I've been away from home and barely listened to them for a few weeks) When I now listen to them properly both with and without different EQ settings, a couple of things strike me.

The sound (without EQ) is characterized by a striking balance between bass, mid-range and treble (But perhaps a perceived elevated treble without a filter. This I suppose is also precisely what is demanded by pros in a mixing situation.) With an filter applied in the treble the sound is soft an leaning towards the ”warm” side. I like this, but this sound signature could probably also be perceived as a problem from an audiophile perspective, because it means the headphone quite brutally reveals flaws in the recordings you listen to.

A general bass elevation by EQ a la Harman may give a more forgiving and initially more "seductive sound", and recordings with weak or poor bass reproduction may sound better with a powerful bass boost. But with a good recording, the bass is absolutely fantastic in the SX5. With EQ, on the contrary, the sound tends to become murkier, I think. The SX5 has a very distinct and contoured bass reproduction, when this can be picked out from the recording.

To the extent that there is a problem with resonances around 250 Hz it isn’t something I managed to discover. I haven't gotten the phone to sound clearer with EQ. It already sounds clear to me, and this clarity is one of it’s strong points. I don't at all recognize my experience with what amirm writes. Someone wrote earlier that I don't need to worry if I think the SX5 sounds good. But it's not concern of that sort that makes me engage in this, but the complete contradiction in, on the one hand, amirm's measurements and listening impressions (and some other people's measurements, apparently) and on the other hand, @solderdude's measurements plus my own and some others listening impressions. I still havent read or heard about anyone that actually have listened to it that don’t like how it sounds. (Maybe amirm is the exception ) How to explain this? I am concidering qestions like How to evaluate headphone measurements, what differentiates measurements, and which recordings are the listening impressions based on. The latter, not least, obviously plays a big role here.

 

[solderdude]

his digital EQ experience is by his own admission decidedly...lacking:

I do use digital EQ by the way, just not when enjoying music as I rarely listen to transducers that actually need it and when they do need a nudge I design and build purpose build analog EQ (active and or passive). For that reason I do not use digital EQ (in general) and if I need EQ it is just very little and analog.

 

[Robbo99999]

I will assume that no matter how I try to explain this you won't accept it because you already KNOW you are right.

Please try to imagine that maybe, just maybe there could be more than FR and distortion when it comes to (electro)acoustics and perception using music vs static test signals that only tell a part of the story... even if it is the largest part.

Acoustic treatment and EQ is not the same thing... period.

You were repeatedly ignoring or repeatedly not reading properly my main points when I was engaging you in discussion, I repeatedly (maybe about 3 times) had to emphasise I was referring to "frequency & distortion" (bolded because you kept coming back to me that I was paying only attention to frequency response), and you never engaged with me in discussion on the points & leading questions I brought up, so it was a non-discussion, so yeah I'm cool, we can be done with the discussion now.

Probably best to just ignore him. When he can't win an argument on a technical front he resorts to snark, ad hominem and appeals to bogus authority (e.g. himself) as we've seen here. In this instance it all stems from his clearly pseudoscientific personal bias against digital EQ, as evidenced in the 'article' he linked previously:

Hilariously, despite lambasting you for not having his extensive hands-on toilet paper experience, his digital EQ experience is by his own admission decidedly...lacking:


Here's what actual professionals and authorities in the field have to say on the topic. First an acoustic engineer who worked/works at Harman (oh and not only a trained listener but a Harman Golden Ear no less), GRAS & Audio Precision:

Is CSD (Cumulative Spectral Decay) really important?

Frequency response is always under suspicion because a lot of audiophiles claim only single one curve can not describe the whole story especially for the time domain. CSD (Cumulative Spectral Decay) contain an extra dimension——time.

www.linkedin.com

(I've only ever seen one headphone that exhibits audible non-minimum phase behaviour shown by significant excess group delay, the Monoprice M1060. It's exceedingly rare.)

Dr. Floyd Toole (my emphasis):



It follows from this last quote that damping would likely in fact result in a more audible resonance than using digital EQ (and potentially even more audible than without, degrading, not improving the stock sound), the latter able to reduce the resonance peak without reducing the Q. Yet another win for EQ vs damping to go along with automatically reducing distortion when bringing down peaks as already mentioned.

And finally, a specific headphone example (source). Here are CSD (waterfall) plots of the HD800 first without, and then with its 6 kHz resonance peak digitally EQed out:
View attachment 302766
View attachment 302767
Oh look, not only has the frequency response ('back' red curve) peak at 6 kHz been removed by this digital EQ, but so has the ringing from it in the time domain. (Of course this all makes sense when it's understood that these domains are mathematically connected via the Fourier transform). No need to take a trip to the toilet for that precious acoustic dumping damping paper after all. Hopefully there'll be no more talking sh*t now that's been cleared up.

Yes, there's some good information there, I'd remembered from somewhere that CSD wasn't important in itself because it was already "explained & dealt with" in frequency response (if frequency response is sorted out then so is CSD), but yes now I can see the specifics now that you're showing that info. Hehe, the "hands on toilet paper experience" line, that was a good one! But yeah, people can use physical inserts in their headphone to change the frequency response, but EQ is better in most ways as long as you've got the ability to apply parametric EQ easily/conveniently in your setup, I don't think I'm changing that viewpoint on the basis of the discussions here over the last 24hrs!

 

[markanini]

I wonder how some people went through life without taking off the pads of their headphones to discover paper surrounding the drivers.

Here's a replacement Beyerdynamic DT770 driver unit with paper on both sides of the driver:

 

[GaryH]

I wonder how some people could go through this whole thread and conclude that some passive (thus digital EQ not available as a design element) headphones having been designed with specifically constructed and tested acoustic dampeners in them means toilet paper is a better tool to fix a headphone's flaws than digital EQ for the consumer.

I can't find anything in the quoted text that specifically and as a prop states that digital EQ must be used, ie applied in the digital domain.

Straw man. The fallacious argument my post and the quotes and data therein refutes is that damping has advantages in the time domain that EQ cannot accomplish. In fact it shows extra damping may actually be audibly detrimental to stock sound by increasing the bandwidth (decreasing the Q-factor) of resonances, thus covering more frequencies and so could be more, not less, likely to be heard.

(The text refers to tests that Floyd Toole did 50 years ago, i.e. around 1970. At that time there were not even digital signal sources like today.)

False. The CSD article references tests done in 1988 (it says Toole has been doing various psychoacoustic tests for 50 years, not that this specific test was done 50 years ago). The additional quotes of Toole in my post are from the 3rd Edition of his Sound Reproduction book, in reference to multiple corroborating experiments done by him and various other acoustic scientists up to 2015. I suggest reading this book so as to actually understand all this.

@GaryH, the link you shared was good read thanks. I got a question if I may please. I understand for transfer function to be a function of frequency, the amplitude response should be linear. But in the case of a resonance, amplitude response is not linear at the resonant frequency, isn't it?

Not sure what you're asking. I think you might be conflating the colloquial meaning of 'linear' (i.e. 'straight', or 'flat' frequency response) here with the very different technical meaning ('transfer function obeys the superposition principle').

 

[bodhi]

I wonder how some people could go through this whole thread and conclude that some passive (thus digital EQ not available as a design element) headphones having been designed with specifically constructed and tested acoustic dampeners in them means toilet paper is a better tool to fix a headphone's flaws than digital EQ for the consumer.

You too? That exact thing kept me up all night.

But anyways, EQ:ing headphones has become the usual high end tweak with questionable benefits. People share 10+ band "recipes" for certain headphones and debate which high Q +/- 1dB corrections should be applied to some better known person's recipe to get just the right sound. Put a few of these tweaked masterpieces to test and start removing some of the bands and you are hard pressed to notice difference let alone decide on preference. That is, if you are really being honest and not falling into "I changed something so it has to improve things".

And sure, EQ is sometimes a must for example with LCD-X which almost seems to be designed to be EQ'd. But even there: LDC-X owners, compare Amir's 5 band EQ with Oratory's 10 band and maybe some tweaked ones from the review thread. Do you notice clear differences?

 

[bodhi]

We are talking about gadgets that are used to listen to music, a non-critical and fully optional aspect of human existence yet look at all this telenovela level drama. One wonders what do people on forums in which they discuss lipid metabolism or cardiovascular health do when they have disagreements, punch each other on the face or hack in to each others' home security systems or something? Or is it just the audio hobby that creates disproportional drama? Anyway.

Let me not tell you about martial arts forums. Especially some sub forums focused on certain styles and different ways to train.

 

[IAtaman]

Not sure what you're asking. I think you might be conflating the colloquial meaning of 'linear' (i.e. 'straight', or 'flat' frequency response) here with the very different technical meaning ('transfer function obeys the superposition principle').

No, I am quite familiar with the non-colloquial meaning of linear, and I did mean in that sense.

All of the arguments laid out in the link you shared relies upon the fact that the transfer function of the loudspeaker is level independent, or at least deviations from linearity of amplitude is mostly small and can be ignored. So if we put in the 1V and get output A, and put 2V and get output B at 3 V we expect to get A+B. But if there is resonance in a certain frequency, that will not be true. One of the defining characteristics of resonance is non-linearity and strong level dependence if I am not mistaken, and as the second part of the article shows, depending on its bandwidth it can be time dependent as well.

So my question to you is whether what I understood and outlined above is true, and if the transfer function is not a function of frequency for the resonant range, does the claim made in the article that you can correct resonances by applying filters still holds true?

 

[IAtaman]

Let me not tell you about martial arts forums. Especially some sub forums focused on certain styles and different ways to train.

Maybe we were all better off not knowing what others thought and believed

 

[Robbo99999]

You too? That exact thing kept me up all night.

But anyways, EQ:ing headphones has become the usual high end tweak with questionable benefits. People share 10+ band "recipes" for certain headphones and debate which high Q +/- 1dB corrections should be applied to some better known person's recipe to get just the right sound. Put a few of these tweaked masterpieces to test and start removing some of the bands and you are hard pressed to notice difference let alone decide on preference. That is, if you are really being honest and not falling into "I changed something so it has to improve things".

And sure, EQ is sometimes a must for example with LCD-X which almost seems to be designed to be EQ'd. But even there: LDC-X owners, compare Amir's 5 band EQ with Oratory's 10 band and maybe some tweaked ones from the review thread. Do you notice clear differences?

You're confusing the practice of EQ'ing an entire headphone from a headphone measurement you find on the internet (and some people's lack of understanding of some of the limitations of that) vs what we're actually talking about in this thread right now which is - insert or EQ? If you know the specific measured effect of an insert on the measured frequency response, then you can replicate that using EQ, and using EQ to replicate that would not be detrimental - therefore it's easier and more precise & testable to use broad ranging (low Q) EQ to tweak your headphones at the upper end rather than painstakingly cutting out sheets of toilet paper, and particularly if you don't have a way to measure (eg flat plate, miniDSP EARS, GRAS, etc) your specific hard mods you're making then you can't be totally sure it's having the desired effect on frequency response that you desire so you're a lot more in the dark when you're experimenting with hard mods if you're a person without a measuring rig, so EQ is the more reliable/testable/easier/without detriment option. Your post is mainly conflating common "EQ mistakes" when EQ'ing an entire headphone (from a headphone measurement you find on the internet) vs instead the specific limited act/effect of using a physical insert (or EQ that replicates that), so your post is mostly not relevant in terms of your criticism of EQ.

(EDIT: by the way there's nothing wrong with 10-band EQ)

 

[GGroch]

This just in: Ollo USC-NX Unit Specific Electronic Calibration add on. Electronic Calibration by WAVES based on the individual measurement files that are made for each S4 S5 headphone.
"It's time for something very cool. We have created an upgrade that will make your OLLO headphones sound flatter than ever.

Out of the box, our headphones are tuned with 1dB SPL tighter tolerance than the 3dB commonly used in room calibration. This ~2dB is the limit we reached on S4 and S5 series using acoustics alone. To achieve even flatter sound, we had to approach hardware and software tuning as a unified whole while respecting different needs and use cases.

That’s why we joined forces with Waves Audio to introduce the USC-NX. It’s a system that builds on the amazing OLLO acoustics and enhances them with a unit specific calibration (USC).... You can buy just the USC upgrade for your specific unit with Waves NX Virtual Room included OR configure your own setup. Head tracking is optional but it makes all the difference. "

YouTube Announcement Here.

 

[solderdude]

They have been working on that for quite while and even have another iron in the fire.

A bit similar in some aspects to Sonarworks (with their individual headphone calibration service) when you buy a headphone from them or sent it in, but with more extra's and different target (also not Harman).
The good thing is that there is no 'heroic' EQ needed and they only have to compensate a little here and there and not have to try to make headphones do things they aren't really designed for or can be made to do well.
Of course the 3D sound emulation requires quite a bit of processing to translate that into stereo and where head-tracking is very helpful.

 

[Robbo99999]

This just in: Ollo USC-NX Unit Specific Electronic Calibration add on. Electronic Calibration by WAVES based on the individual measurement files that are made for each S4 S5 headphone.
"It's time for something very cool. We have created an upgrade that will make your OLLO headphones sound flatter than ever.

Out of the box, our headphones are tuned with 1dB SPL tighter tolerance than the 3dB commonly used in room calibration. This ~2dB is the limit we reached on S4 and S5 series using acoustics alone. To achieve even flatter sound, we had to approach hardware and software tuning as a unified whole while respecting different needs and use cases.

That’s why we joined forces with Waves Audio to introduce the USC-NX. It’s a system that builds on the amazing OLLO acoustics and enhances them with a unit specific calibration (USC).... You can buy just the USC upgrade for your specific unit with Waves NX Virtual Room included OR configure your own setup. Head tracking is optional but it makes all the difference. "

YouTube Announcement Here.

It's good that they're basically removing the element of unit to unit variation when you buy the headphone, but of course the target curve is only as good as the target curve and same is true for the Virtual Room software stuff. I like that they're effectively removing unit to unit variation though.

 

[Zensō]

The USC-NX seems to be designed to compete with Slate Digital. Personally, I’m very suspect of any room emulation software (not to be confused with cross feed), which always colors the sound to a degree that it may negatively affect mix decisions. The more accurate unit-to-unit calibration is a nice feature, though if they were already at 2 dB as stated, the additional 1 dB is not likely to result in any real world improvement. Ultimately, the only advantage seems to be for those who are already using Waves NX room sims.

 

[solderdude]

1dB in a wide Q is audible.
But here again... it is based on measurements made on 2 different standards which both are not 'reality'.
We would really have to wait what the actual results will be and if an average EQ will emerge and what that will do to the the average S4X, S4R and S5X.

I know Rok is working a lot with studios doing multichannel recordings and the S5X tuning is based on what the found (including the somewhat elevated treble) which is needed to get the simulation to be 'correct'.

Let's see where this takes them. I have my concerns.